Supply chain scheduling with delivery costs

Abstract

<p> Supply chain management has been one of the most important issues in manufacturing industries. In order to improve customer satisfaction, suppliers usually extend their inbound production operations to outbound logistics operations. To improve the overall operational performance, it is necessary to study scheduling models which consider inbound production and outbound deliveries simultaneously. This thesis deals with supply chain problems on the operational level using deterministic models. </p> <p> Meeting due dates is always one of the most important concerns in scheduling and supply chain management. In most supplying contracts, customers require that suppliers either meet contracted due dates or pay tardiness penalties. In order to save delivery costs, suppliers usually deliver jobs in batches. Therefore, we will study supply chain scheduling problems with delivery costs, where our goal is to minimize the sum of the weighted number of tardy jobs and the batch-delivery costs on a single machine. </p> <p> In traditional manufacturing system, due dates are not considered as given by exogenous decisions. In modern supply chains, however, due dates are determined by taking into account the system's ability to meet the assigned due dates, which can be quoted with certain costs. Therefore, we will study supply chain scheduling problems with delivery costs and due date assignment, where our goal is to minimize the sum of the weighted number of tardy jobs, the due-date-assignment costs and the batch-delivery costs on a single machine. </p> <p> As we know, most machine scheduling models are intractable in terms of computational complexity. Therefore, for our above problems, which are even harder, we first prove their computational complexity. Then we propose pseudo-polynomial algorithms for optimal solutions. For some problems, the pseudo-polynomial algorithms perform in polynomial time for some special cases. Finally, we develop efficient approximation algorithms or fully polynomial time approximation schemes, which can be implemented easily in practice. </p>